812
Inorg. Chem. 1994, 33, 8 12-8 15
Tungsten(6+) Tris(pinaco1ate): Structure and Comments on the Preference for an Octahedral for a do Complex in the Presence of Strong Geometry Relative to Trigonal Prismatic T-Donor Ligands
(ab)
M. H. Chisholm,' Ivan P. Parkin, William E. Streib, and 0. Eisenstein' Department of Chemistry and Molecular Structure Center, Indiana University, Bloomington, Indiana 47405, and Laboratoire d e Chimie ThCorique, Batiment 490, UniversitC d e Paris-Sud, 91405 Orsay, France Received September 29, 1993'
The reaction between W(NMe& and HOCMeZCMezOH (pinacol, 3 equiv) in ether/hexanes yields the pale yellow compound W(OCMezCMe20)3(1) as pale yellow cubic crystals. Crystal data at -155 OC: a = 16.841(2) A, b = 9.878(1) A, c = 13.373(1) A, p = 109.60(1)O, Z = 4, d,l, = 1.69 g ~ m - space ~ , group C2/c. The molecule has a crystallographically imposed Cz axis of symmetry but may basically be described as having a distorted octahedral W 0 6 core, W-0 = 1.92( 1) 8, (average). Variable-temperature 'H N M R data reveal two Me proton signals (+80 to -80 OC) in tOlUene-ds and thus indicate that enantiomerization via a trigonal prismatic intermediate is slow on the N M R time scale. The preference for the do-octahedral WO6 geometry is explained on the basis of oxygen P, to W d, bonding, and these results are compared with those for other do-ML6 complexes and do-M(che1ating ligand)3 complexes.
Introduction Most of the hexacoordinated transition metal complexes are octahedral. This is in particular the case of d6 complexes where angular distortion away from ideal 90' angles are due to the steric demands of thecoordinated ligands. Lower electron counts at the metal result in a higher variety of structures: octahedral, trigonal prismatic notably in the case of do metal centers,l-8 and bicapped tetrahedra in the case of d4 metal ~ e n t e r s . l - ~ The - l ~ do systems have been especially controversial. VSEPR rules would give a preference for an octahedral geometry. The original proposal,I,Zbased on a second-order Jahn-Teller argument that do-MR6 ( R = pure u-donor ligands) may not have an octahedral geometry has been challenged by state of the art ab initio calculations,3 and contradictory results have been obtained. A consensus has been reached on do-MH6. The equilibrium geometry has been calculated to be of either 0 3 1 or C3" symmetry for different metal center^.^^^ Experimental data have supported these theoretical results since WMe6 has been found to have D3h symmetry by electron diffraction.*" Recent theoretical studies have confirmed that MH6 and MMQ have similar structural properties.sb Replacing pure u-donor ligands by 7r-donor ligands decreases the tendency for distortion away from 01 symmetry since it increases the HOMO-LUMO gap. Contradictory results have appeared in the case Of CrF6 probably because F is a moderate
* To whom correspondence should be addressed: M.H.C., Indiana University; O.E., Universite de Paris-Sud. 0 Abstract published in Aduance ACS Abstracts, January 15, 1994. (1) Hoffmann, R.; Howell, J. M.; Rossi, A. R. J . Am. Chem. SOC.1976,98, 2484. (2) Demolliens, A.; Jean, Y.; Eisenstein, 0.Organometallics 1986,5, 1457. (3) Cameron, A. D.; Fitzgerald, G.; Zerner, M. C. Inorg. Chem. 1988, 27, 3437. (4) Kang, S. K.; Albright, T. A.; Eisenstein. 0. Inorg. Chem. 1989, 28, 1611. (5) Marsden, C. J.; Wolynec, P. P. Inorg. Chem. 1991, 30, 1681. ( 6 ) Jonas, V.; Frenking, G.; Gauss, J. Chem. Phys. Lett. 1992, 194, 109. (7) Hope, E. G.;Levason, W.; Ogden, J. S.Inorg. Chem. 1991,30, 4874. (8) (a) Haaland, A.; Hammel, A.; Rypdal, K.; Volden, H. V. J . Am. Chem. SOC.1990,112,4547. (b) Kang,S. K.; Tang, H.;Albright,T.A. J . Am. Chem. SOC.1993, 115, 1971. (9) Kubacek, P.; Hoffmann, R. J . Am. Chem. SOC.1981, 103, 4320. (10) Templeton, J. L.; Winston, P. B.; Ward, B. C. J . Am. Chem. SOC.1981, 103,7713. (1 1) Kamata, M.; Hirotsu, K.; Higuchi, T.; Tatsumi, K.; Hoffmann, R.; Yoshida, T.; Otsuka, S. J . Am. Chem. SOC.1981, 103, 5772.
0020-166919411333-0812$04.50/0
7r d o n ~ r . ~It, ~is probable that the potential energy surface describing the structural distortion away from 01 symmetry is rather flat for such species. Experimental data on the elusive CrF6 complex are in support of an octahedral structure, but this matter continues to be discussed extensively in the literature.12-l4 Better 7r donors than F should certainly increase the preference for oh symmetry. We report here the synthesis and characterization of tungsten(6+) tris(pinacolate), W(OCMe2CMe20)3 (l),and show that, despite the presence of the three chelating ligands, the barrier to enantiomerization via a trigonal prismatic D3h geometry is slow on theNMR time scale. The preference for octahedral geometry, albeit distorted, is understood in terms of oxygen pz to tungsten d, bonding.
Results and Discussion Synthesis. The reaction between 3 equiv of pinacol and tungsten hexakis(dimethy1amide) in diethyl ether/hexane produces tungsten tris(pinaco1ate) in high yield (eq 1). Monitoring the reaction W(NMe,),
+ 3HOCMe,CMe,OH
-
W(OCMe,CMe,),
+ 6HNMe,
(1)
by IH N M R spectroscopy showed the production of HNMez but revealed no detectable tungsten-containing intermediates. The substitution reactions of W(NMe2)6 with alcohols have been studied, and complexes of formulas W(OMe)6 and W(OEt)6 have been synthesized.15 For larger alcohols such as t-BuOH and neopentanol, no reaction with W(NMe2)6 occurs at room temperature. Presumably, steric hindrance prevents alcoholysis. Thus it is noteworthy that pinacol, which has essentially the same stericshielding at the a-carbon as t-BuOH, undergoes substitution reactions with W(NMe2)6. The pK, of pinacol is smaller (more (12) Pierlott, K.; Rws, B. 0. Inorg. Chem. 1992, 31, 5353. ( 1 3) Neuhaus, A.; Frenking, G.; Huber, C.; Gauss, J. Inorg. Chem. 1992,31, 5355. (14) Jacobs, J.; Miiller, H. S. P.; Willner, H.; Jacob, E.; Bdrger, H. Inorg. Chem. 1992, 31, 5357. (15) Chisholm, M. H.; Extine, M. W.; Stager, M. W. Inorg. Chem. 1977, 16, 1794.
0 1994 American Chemical Society
Tungsten(6+) Tris(pinaco1ate) acidic) than that of t-BuOH,I6 and probably the chelate effect of the diolate ligand makes subsequent coordination more accessible. Characterization. Tungsten tris(pinaco1ate) was characterized by VT IH N M R , IR, melting point, mass spectrometry, microanalysis, and X-ray crystallography. The compound crystallizes as pale yellow cubes by cooling of a saturated diethyl ether/hexane solution. It is air stable both in solution and in the solid state (for over a year), soluble in aromatic and aliphatic hydrocarbon solvents, and unreactive toward alcoholysis by MeOH and EtOH. On heating to 182 ‘C, it decomposes. However, an electron impact mass spectrum did detect the parent ion at 534 amu with the correct isotopic abundances for an included tungsten atom. Other fragmentation peaks occur at M+ C(CH&H, M+ - OC(CH3)2H, and M+ - C2Me402,all with the correct tungsten isotopic correlations. The IR spectrum of W(02C2Me& shows the expected pinacolate vibrations. Solution-State lH NMR. The IH N M R of W(02C2Me4)3 is temperature invariant in C ~ D S C D from ~ -90 to +I10 ‘c and shows two resonances of equal intensity at 6 = 1.54, 1.12 ppm. This implies that there are two different environments for the methyl groups in W(02C2Me4)3. The solid-state structure of W(OtC2Me4)3 shows one C2 axis and one pseudo-C3 axis. In solution the five-membered rings will be expected to be undergoing rapid ring flipping even a t -90 ‘C on the N M R time scale. The rapid interconversion of the C2 axis by ring flipping generates on the N M R time scale effectively three C2 axes, and this makes all three rings equivalent. However, this process does not make the four methyl groups on each pinacolate ligand equivalent. For each M(che1ating ligand)3 enantiomer, with time-averaged 0 3 symmetry, there will be two sets of Me ligands. However, if the W06 core were stereochemically labile with respect to enantiomerization, i.e. attainment of a trigonal prismatic geometry, then the combined ring flipping and WO6 skeletal rearrangements would lead to only one time-averaged Me signal. The barrier for enantiomerization involving a trigonal prismatic geometry must then be greater than 15 kcal mol-’. Of course, a fluxional process based on W-0 bond cleavage might have been possible-this too is ruled out by the N M R behavior and is presumably less likely for a W6+-containing compound with three dianionic chelates than, e.g., for a M3+(chelatingligand)3 complex where the ligand has a uninegative charge, as for example in M(acac)3 complexes where M = Al, In, Ga and acac is the anion derived from deprotonation of 2,5-pentanedione.17 Crystal and Molecular Structure of W(02C2Me& (1). In the space group C2/c there is one unique molecule of W(02C2Me4)3 in the unit cell. Two views of the molecule are given in Figure 1, and selected bond lengths and angles are given in Table 1. A summary of the crystal data is given in Table 2. The compound has a distorted octahedral W 0 6 moiety and a crystallographically imposed 2-fold axis of symmetry bisecting O(lo), O(lo)’, and W(1). The structure is very similar to that adopted by w (02C2H4) 3. The crystal structure of W(02C2Me4)3 is consistent with do W(6+) being surrounded by six uninegative 0-R ligands. All the Wadistances are, within the 3ucriteria, the same in W(OzC2Me4)3, at 1.916(3) A (average). This distance is comparable to those observed in W2(OR)6complexes at 1.88 (average)19 and is consistent with some M d,-0 pr interaction. For W(02C2Me& the twist angle 8, as calculated from the centroids viewed down the pseudo-C3 axis, is 34’. This indicates (16) pKa Prediction for Organic Acids and Bases; Perring, D. D., Demprey, B., Serjent, P., Eds.; Chapman and Hall: London, 1981. (1 7) For a discussion of the dynamic NMR behavior of M(che1ating ligand), complexes, see: Holm, R. H. In Dynamic NMR, Cotton, F . A., Jackman, L. M., Eds.; Academic Press: New York, 1975; Chapter 9. (18) Scherle, V. J.; SchrMer, F. A. Acta Crystallogr. 1974, B20, 2772. (19) Chisholm, M. H. Polyhedron 1983, 2, 681.
Inorganic Chemistry, Vol. 33, No. 4, 1994 813
Figure 1. ORTEP drawing of the W(02C2Me4)s molecule showing the atom-numbering scheme (top) and a view of the molecule viewed down the crystallographically imposed C2 axis of symmetry.
Table 1. Selected Bond Distances (A) and Angles (deg) for W(02C2Meds (1) Distances 1.914(3) 0(2)-C(3) 1.915(3) 0(5)-C(4) O(lo)
